I have this query: data.stackexchange.com/stackoverflow/query/1628473#graph where I probably abuse log10 to get a decent graph. But I use that outcome to make a point here: meta.stackexchange.com/a/381364

I'm wondering if there is a better suited method / statistical function to show or analyze that bump for "users in the log10 = 1" group.

Just grouping on reputation and then count makes a mess: data.stackexchange.com/stackoverflow/query/1628528#graph